Every run: the agent's full trajectory, its submissions over the 20-hour budget, and the final score.
| # | Model | Reward | Cost | In tok | Out tok | Time | Steps | Submits |
|---|---|---|---|---|---|---|---|---|
| 1 | GPT-6 Astra | 0.9700 | $116.67 | 60.2M | 205k | 18h 38m | 255 | 8 |
| 2 | GPT-6 Astra | 0.9160 | $256.74 | 124.3M | 279k | 19h 10m | 365 | 11 |
| 3 | GPT-6 Astra | 0.9000 | $49.97 | 27.6M | 160k | 19h 17m | 158 | 6 |
| 4 | GPT-6 Astra | 0.8840 | $265.88 | 128.3M | 310k | 19h 14m | 390 | 8 |
| 5 | GPT-6 Astra | 0.6760 | $444.71 | 214.4M | 392k | 19h 40m | 521 | 10 |